General Motors announced this week that it had struck a deal with the United Auto Workers union to cut $3 billion a year from its spending on health benefits for retirees. Even this huge sum may be too little to secure GM's future. The company has plenty of things to worry about besides its spending on health: high gas prices and falling demand for its most profitable lines, to name two. Still, without the new agreement, the company's prospects would have been even worse.

GM is not alone. Many American businesses complain that they are being crushed by health care costs. Competitors abroad, they emphasize, are spared this burden, so America Inc. is at a grave disadvantage. It is not a new complaint, but the persistent trade deficit and the perilous finances of iconic industrial giants like GM give it a new edge.

Is it really true, though, that corporate health costs put American industry at an international disadvantage, as GM and many other companies say? The idea seems plausible, but it is more slippery than you might think.

In principle, as far as current workers are concerned, health benefits are just another form of pay. If companies pay each of their workers $1,000 a year in health benefits (and suppose, for the sake of argument, that the workers would have wished to buy them in any case), then it ought to be possible to pay them $1,000 a year less in wages and other forms of compensation, and still be about as successful as before in recruiting and retaining employees. Total pay is the same in each case.

To judge international competitiveness, labor costs must be seen as a whole and must be measured relative to productivity. Overall, American workers are productive enough to be competitive at prevailing rates of total compensation; otherwise, the rate of unemployment would be much higher. (What about the trade deficit? Doesn't that prove that lack of competitiveness is a problem? No. The deficit has little to do with competitiveness. Mainly, it reflects the strength of demand in the United States relative to the strength of demand in Europe and Asia.)

Now, it is true that several things complicate the GM picture. One is "legacy costs"—benefits due to workers who have retired. GM made expensive long-range promises to its retirees, on pensions as well as on health care, which the firm can no longer afford and wants to undo. Whether those promises were ever prudent is doubtful, and the company's provision for its future liabilities was evidently inadequate. A good rule is not to make promises you cannot keep. A related complication is labor relations. GM may be shrewd to complain about health care costs rather than wages or labor costs more broadly, even if they all amount to the same thing, so long as the UAW is more willing (and it probably is) to swallow a cut in health benefits than a cut in wages or jobs.

And politics comes in as well, of course. GM and other manufacturers may believe (and they are probably right about this also) that if they complain long enough and loudly enough about health and pension costs, the government will give them some kind of relief at taxpayers' expense. Complaining about high wages—even though, again, the issue is no different in principle—would be less likely to attract additional tax breaks or other forms of subsidy.

They crucial word in that last sentence, though, is "additional"—because the deals that businesses and their workers come to over wages, benefits, and hours on the job are already twisted badly out of shape by tax policy. If it were not for tax policy, and the big incentive that it gives employers to pay their workers in tax-sheltered benefits rather than in cash, GM would most likely not be providing health benefits in the first place.

This is the most important complication of all. The tax preference for employer-provided health care has a big unintended consequence: It makes health care more expensive.

How? By pushing decisions about health care—about the trade-off between costs and benefits—away from the actual consumers of health care. As a result, end users have little reason to economize. Managed care and insurance-policy deductibles mitigate the problem but do not resolve it. At the margin, with somebody else (apparently) paying the insurance premium, patients get additional health services at a fraction of the cost. So far as economic pressures are concerned, demand is unconstrained.

To be sure, there are other, non-economic, constraints: People do not demand surgery or chemotherapy for the fun of it. But they may very well demand the most expensive drugs, the most expensive procedures, the most expensive ancillary services, with no regard at all for the balance between efficacy and cost—a balance that they would certainly strike if they were paying out of their own pockets. In the aggregate, users still pay for those services—through either taxes or lower wages, or both. At the margin, they behave as though they don't. The result, predictably, is rapid inflation in the cost of care.

So this tax preference, though it masquerades as pro-worker, is really anti-worker. The idea that the firm, not the worker, is paying for employer-provided health care is an illusion—similar to the idea that the gift you get when you shop at a particular store or fly a particular airline is "free." Health benefits—to repeat—are just another form of compensation, and total compensation, in most industries, is set by the market. If health care costs more because of the policy—and it does—in the end, it is workers who will pay. One way or another, their purchasing power will be lower. The illusion that workers get health care for nothing costs them a lot.

By the way, even this genuine burden of additional cost has little or no bearing, in the end, on the economy's competitiveness. Wages and other benefits adjust to cancel out the effect on total labor costs. Competitiveness is not the point. GM is correct, nonetheless, to say that its spending on health care, like every other company's, is more than it should be.

The larger issue is not competitiveness, but economy-wide living standards. Never mind who pays, or seems to be paying, for health care; never mind whether the cost is met out of taxes, out of pre-tax wages, or out of household disposable income. If health care across the economy costs more than it should, people have less to spend on other things, and they are worse off overall. That is the problem.

"Healthy Competition," a new study by Michael F. Cannon and Michael D. Tanner, just published by the Cato Institute, makes the interesting point that in one way—as measured by the proportion of health spending channeled through third parties—America's health care system is more thoroughly "socialized" than many in Europe, even more than Canada's. Despite America's so-called market in health care, the user is no closer to the system's costs than are users in countries that have opted for outright nationalization.

The big difference lies elsewhere—in the degree to which top-down rationing is used to contain costs. In single-payer universal-coverage systems such as Britain's National Health Service, the rationing is brutal. It results in dirty hospitals and months-long waiting lists for urgently needed treatments. America's spare-no-expense approach gives superlative results for the rich, excellent standards for most Americans, and service even worse than in the NHS for many of the uninsured—at a total cost that is likely in the end to be politically unsupportable. According to the Congressional Budget Office, at present trends, Medicare alone will likely cost the United States a larger share of GDP in the 2030s than the entire NHS costs Britain today. Neither model, in short, can be called a success.

Cannon and Tanner rightly emphasize that if greater outlays were buying correspondingly better results, the country would have no reason to complain. The results in this country are indeed better—but the price is higher than necessary. As they argue, and document, in a system substantially without pressure from users to curb costs, a great deal of spending is wasted.

It is often said that health care is the only sector of the economy where advancing technology drives costs up rather than down. Up to a point, that is fine. If expensive new procedures and medicines can significantly prolong life and improve its quality, people will want to buy. But technology properly deployed could push costs down at the same time. Health care need not be an exception to the rule. Remote monitoring and supervision of patients with chronic illnesses, and less time spent in hospital, is one avenue beginning to be explored. There are many others. Cost-reducing innovations of many different kinds would come much faster if users were pressing for them. The catch is, they will press for them only when they have a good economic reason to.